Agent-skills kibana-streams

Name: kibana-streams
Author: elastic

install

source · Clone the upstream repo

git clone https://github.com/elastic/agent-skills

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/elastic/agent-skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/kibana/streams" ~/.claude/skills/elastic-agent-skills-kibana-streams && rm -rf "$T"

manifest: skills/kibana/streams/SKILL.md

source content

Kibana Streams

Read stream metadata, settings, queries, significant events, and attachments, and manage stream lifecycle (enable, disable, resync) via the Kibana Streams REST API. Streams are an experimental way to manage data in Kibana — expect API and behavior changes. This skill covers read operations and lifecycle only; create, update, delete, fork, and other mutating operations may be added in a later version.

For detailed endpoints and parameters, see references/streams-api-reference.md.

When to use

Listing all streams or getting a single stream's definition and metadata
Reading a stream's ingest or query settings
Listing a stream's queries
Reading significant events for a stream
Listing attachments (dashboards, rules, SLOs) linked to a stream
Enabling, disabling, or resyncing streams

Prerequisites

Item Description

Kibana URL

Kibana endpoint (e.g.

https://localhost:5601

or a Cloud deployment URL)

Authentication API key or basic auth (see the elasticsearch-authn skill)

Privileges

read_stream

for read operations;

manage_stream

for lifecycle APIs

Use the space-scoped path

/s/{space_id}/api/streams

when operating in a non-default space. For role configuration (Kibana feature privileges and Elasticsearch-level permissions), refer to Streams required permissions.

API base and headers

Base path:

GET

POST

<kibana_url>/api/streams

(or

/s/<space_id>/api/streams

for a space).

Read operations: Typically do not require extra headers; follow the official API docs for each endpoint.
Lifecycle operations:
```
POST /api/streams/_disable
```
,
```
_enable
```
, and
```
_resync
```
are mutating — send
```
kbn-xsrf: true
```
(or equivalent) as required by your Kibana version.

Operations (read + lifecycle)

Read

Operation	Method	Path
Get stream list	GET	`/api/streams`
Get a stream	GET	`/api/streams/{name}`
Get ingest stream settings	GET	`/api/streams/{name}/_ingest`
Get query stream settings	GET	`/api/streams/{name}/_query`
Get stream queries	GET	`/api/streams/{name}/queries`
Read significant events	GET	`/api/streams/{name}/significant_events`
Get stream attachments	GET	`/api/streams/{streamName}/attachments`

Lifecycle

Operation	Method	Path
Disable streams	POST	`/api/streams/_disable`
Enable streams	POST	`/api/streams/_enable`
Resync streams	POST	`/api/streams/_resync`

Path parameters:

{name}

and

{streamName}

are the stream identifier (same value; the API docs use both names).

Lifecycle and retention (ingest settings)

Ingest settings (

GET /api/streams/{name}/_ingest

) expose two separate lifecycle areas:

Stream lifecycle (
```
ingest.lifecycle
```
) — Controls how long the stream's data is retained. Use
```
lifecycle.dsl.data_retention
```
(e.g.
```
"30d"
```
) for explicit retention, or
```
lifecycle.inherit
```
for child streams. This is what users usually mean when they ask to "set retention", "update retention", or "change the stream's retention".
Failure store lifecycle (
```
ingest.failure_store.lifecycle
```
) — Controls retention of failed documents only (documents that did not process successfully). Users rarely need to change this unless they explicitly mention the failure store or failed-document retention.

When a user asks to set or update retention, target the stream's main lifecycle (

lifecycle.dsl.data_retention

), not the failure store, unless they specifically ask about failure store or failed documents.

Examples

List streams

curl -X GET "${KIBANA_URL}/api/streams" \
  -H "Authorization: ApiKey <base64-api-key>"

Get a single stream

curl -X GET "${KIBANA_URL}/api/streams/my-stream" \
  -H "Authorization: ApiKey <base64-api-key>"

Get stream queries

curl -X GET "${KIBANA_URL}/api/streams/my-stream/queries" \
  -H "Authorization: ApiKey <base64-api-key>"

Get significant events or attachments

# Significant events
curl -X GET "${KIBANA_URL}/api/streams/my-stream/significant_events" \
  -H "Authorization: ApiKey <base64-api-key>"

# Attachments (dashboards, rules, SLOs linked to the stream)
curl -X GET "${KIBANA_URL}/api/streams/my-stream/attachments" \
  -H "Authorization: ApiKey <base64-api-key>"

Disable, enable, or resync streams

# Disable streams (request body per API docs) — deletes wired stream data; warn and confirm before proceeding
curl -X POST "${KIBANA_URL}/api/streams/_disable" \
  -H "Authorization: ApiKey <base64-api-key>" \
  -H "kbn-xsrf: true" \
  -H "Content-Type: application/json" \
  -d '{}'

# Enable streams
curl -X POST "${KIBANA_URL}/api/streams/_enable" \
  -H "Authorization: ApiKey <base64-api-key>" \
  -H "kbn-xsrf: true" \
  -H "Content-Type: application/json" \
  -d '{}'

# Resync streams
curl -X POST "${KIBANA_URL}/api/streams/_resync" \
  -H "Authorization: ApiKey <base64-api-key>" \
  -H "kbn-xsrf: true" \
  -H "Content-Type: application/json" \
  -d '{}'

Check the Streams API operation pages for request/response bodies (e.g. request body for _disable/_enable/_resync if required).

Guidelines

When the user asks to set or update retention, assume they mean the stream's data retention (
```
ingest.lifecycle
```
/
```
lifecycle.dsl.data_retention
```
). Do not change only the failure store retention unless they explicitly ask about the failure store or failed documents.
Other mutating operations (create, update, delete, fork, bulk query management, attachment management, and more) are not supported by this skill. See references/streams-api-reference.md for the full list of deferred operations.
Disabling streams can lead to data loss for wired streams. The disable API deletes wired stream data (classic stream data is preserved). Before calling disable, warn the user and confirm they understand the risk (and have backed up or no longer need the data).
Prefer read operations when the user only needs to inspect stream state; use lifecycle APIs when they need to enable, disable, or resync streams.